Hard K-means and responsibilities If you did not read the first part of the clustering series. Please go check it out. I use the same data points and this post starts from troubleshooting the hard K-means algorithm in the previous post.
In the previous post, we defined assignment. The equivalent representation of this assignment of points to clusters is given by responsibilities, $r^{(n)}_k$. In the assignment step, we set $r^{(n)}_k$ to one if mean k is the closest mean to datapoint $ {\textbf x}^{(n)}$; otherwise, $r^{(n)}_k$ is zero.